 |
 |
We used the INCstar method (RIA) until late 1996 when we switched to the Dade Behring method (EMIT). The INCstar method supplied serum standards for a whole blood assay which lead us to investigate making our own standards. We obtained the cyclosporin from Sandoz and spiked the plasma fraction of whole blood with ethanolic solutions of cyclosporin. We first filled a volumetric with whole blood, allowed the cells to settle and removed a quantity of plasma to spike which was then replaced into the volumetric to obtain whole blood standards. This took a bit of nifty maths (well for my brain anyway they were nifty) to get the correct concentrations but it worked very well. We would then split each standard, run one as whole blood and the other after a freeze thaw cycle to haemolyse and compare the standard curves obtained. In most cases the curves would be identical and we felt we had valid standards traceable to a the sample matrix we were using. We stoped this practice when we switched to the EMIT method as Haemolysed standards were supplied but I have wondered whether the commercial standards have been prepared so rigourously. Also we found spiking haemolysed blood produced very variable recoveries, what are others experience? My real concern with cyclosporin is the QC of the assay. INCstar supplied serum based controls and Dade Behring recommend lyophilsed haemolysed whole blood controls but do these realy represent adequate QC of the extraction phase a whole blood samples, certainly the extracted "pellet" in whole blood and haemolysed samples are visably different. We now reluctantly don't produce our own inhouse controls but use Bio-Rads. Do any labs have any "neat" answers do this problem? The question of QCing an assay with a matrix different from the patient sample matrix is an old problem but should not commercial material have to show equivalent performance in terms of imprecision etc, before it can be marketed? Alun Price Clinical Chemistry Northern General Hospital Herries Road Sheffield S5 7AU 0114 2715321 [log in to unmask]